a thoughtful web.
Good ideas and conversation. No ads, no tracking.   Login or Take a Tour!
comment by ooli
ooli  ·  3184 days ago  ·  link  ·    ·  parent  ·  post: Lee Se-dol beats AlphaGo in game 4

The unreported story being that the AI purposely lose the match. It's the one thing we should expect from Multivac





user-inactivated  ·  3183 days ago  ·  link  ·  

Is this true? I don't keep track of the news, and perhaps this is why I'm buffled by this idea - that it would want to lose on purpose.

ooli  ·  3183 days ago  ·  link  ·  

No it's a joke.

Isaac Asimov wrote some sci-fi book about an AI named Multivac. Most of the time Multivac made just enough mistake so human can feel a bit less useless. The AI only want their happiness. Being less than perfect help for that purpose.

It would be nice that AlphaGo could feel the 'need' to lose for the beauty of this game about intuition and harmony.

user-inactivated  ·  3183 days ago  ·  link  ·  

Thanks for clarifying that.

    It would be nice that AlphaGo could feel the 'need' to lose for the beauty of this game about intuition and harmony.

Otherwise it would ruin what we consider the beauty of the unknown and shred the fog of mystery, wouldn't it? The fact that we can program a machine to win a game of intuition - implying a sort of system behind it - that is.